Review Request: fix border calculation by positioning panels based on their location

[Aaron Seigo]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://mattr.info/r/238/#review229
-----------------------------------------------------------


the problem is how we put panels on the canvas. right now we sit them atop the screen they are associated with.

i think what we should do is stack all horizontal panels with x == 0 (for flush left panels, anyways; x + offset for non-left-flush panels) and all vertical panels starting at, say, and x == 100000 (or whatever, just big enough to never be run into by horizontal panels).

subsequent horizontal panels would have increasingly negative y() (which is irrelevant for their use and geometry calcs) and verticals would have increasingly positive x(), iow stacking.

then we only need to use the width/height of the screen rect for horiz/vert panels, much as you have done here.


/trunk/KDE/kdebase/workspace/plasma/containments/panel/panel.cpp
<http://mattr.info/r/238/#comment177>

    the point of using the screenRect is to take into consideration multiple screens. we place containments above the screen they are associated with. so this line breaks with two screens and a panel on the second screen.



/trunk/KDE/kdebase/workspace/plasma/containments/panel/panel.cpp
<http://mattr.info/r/238/#comment178>

    ditto on multi-screen problems.


- Aaron


On 2008-03-01 22:42:31, Chani Armitage wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://mattr.info/r/238/
> -----------------------------------------------------------
> 
> (Updated 2008-03-01 22:42:31)
> 
> 
> Review request for Plasma.
> 
> 
> Summary
> -------
> 
> 
> Diffs
> -----
> 
>   /trunk/KDE/kdebase/workspace/plasma/containments/panel/panel.cpp
> 
> Diff: http://mattr.info/r/238/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Chani
> 
>